Mineral Processing and Extractive Metallurgy Review

ISSN: 0882-7508 (Print) 1547-7401 (Online) Journal homepage: https://www.tandfonline.com/loi/gmpr20

DIRECT REDUCTION OF MIXED MAGNETITE AND

COAL PELLETS USING INDUCTION HEATING

REUBENDRAN CHELLAN , JONATHAN POCOCK & DAVID ARNOLD

To cite this article: REUBENDRAN CHELLAN , JONATHAN POCOCK & DAVID ARNOLD
(2004) DIRECT REDUCTION OF MIXED MAGNETITE AND COAL PELLETS USING
INDUCTION HEATING, Mineral Processing and Extractive Metallurgy Review, 26:1, 63-76, DOI:
10.1080/08827500490477612

To link to this article: https://doi.org/10.1080/08827500490477612

Published online: 31 Aug 2006.

Submit your article to this journal

Article views: 159

Citing articles: 3 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=gmpr20
Mineral Processing & Extractive Metall. Rev., 26: 63
76, 2005
Copyright # Taylor & Francis Inc.
ISSN: 0882-7508 print
DOI: 10.1080=08827500490477612

DIRECT REDUCTION OF MIXED MAGNETITE AND

COAL PELLETS USING INDUCTION HEATING

REUBENDRAN CHELLAN
JONATHAN POCOCK
DAVID ARNOLD
School of Chemical Engineering, University of KwaZulu-Natal,
Durban, South Africa

A low-temperature process for iron-making, using mixed pellets of magnetite

and domestic coal with an organic binder was investigated. Reduction tests
performed in an induction furnace resulted in a product direct reduced iron
(DRI) containing 75
80% metallic iron. The coal energy consumption of
23.71 GJ=ton DRI compares favorably with current DRI processes, although
the total energy consumption of 33.65 GJ=ton is relatively high for iron-
making. The use of low-grade coal as a reductant may improve the economics
of the process as high-grade coking coal supplies dwindle worldwide.

Keywords: iron-making, direct reduction, DRI, induction heating

The production of pig iron in the blast furnace ranks foremost among all
the iron-making processes, constituting 59% of the world iron production
in 2001 (Moore 2001). Coke is used in large quantities in iron blast fur-
naces to promote a combination of direct and indirect reduction within
the furnace. Direct reduction involves the carbon reacting directly with the

We would like the thank Lignotech South Africa for sponsorship, the Geology Depart-
ments of the Westville and Howard College campuses, University of KwaZulu-Natal, for
respective X-Ray diffraction analysis and the SEM unit of Howard College campus for
analysis performed.
Address correspondence to Jonathan Pocock, School of Chemical Engineering,
Howard College Campus, University of KwaZulu-Natal, King George V Avenue, Durban
4041, South Africa. E-mail: Pocockj@nu.ac.za

63
64 R. CHELLAN ET AL.

iron oxide to form iron, while indirect reduction involves the carbon
monoxide, which is a product of direct reduction, becoming the reducing
agent to form metallic iron. The product is a liquid ‘‘pig iron,’’ containing
up to 5% carbon and 1% silicon (Peacey and Davenport 1979).
The blast furnace process faces many problems, such as low flex-
ibility in production capacity (Sasaki 1998), emission of air pollutants
(Lungen 2001), and high investment cost (Gielen 2002).
The current major problem, however, is the decrease in supplies of
coking coal. The coal required for coke production must have specific
properties and production of coke is not environmentally friendly, as
volatiles have to be burnt off. In the world, there are limited reserves of
coking coals (Yusfin and Pashkov 2002). Research conducted by the
Department of Mechanical Engineering, University of Cape Town, South
Africa (Anon 2001) confirms that only approximately 2% of the large coal
reserves in South Africa, i.e., 55 billion tons, are suitable for coke making.
Due to the lack of good coking coal within South Africa and
dwindling supplies worldwide, new iron-making processes are being
developed to use cheaper coal supplies. Processes that produce iron by
reduction of iron oxide below the melting point of the metal produced are
generally classified as direct reduction processes and the products are
referred to as direct reduced iron (DRI) (Feinman 1999). DRI is the
product that has a high metallization degree, i.e., >90% (Els 2001),
obtained by reducing iron bearing materials with natural gas or non-
coking coal in a reducing unit. The product, which is porous, is also
known as sponge iron. In 2001, the worldwide production of DRI
amounted to 39 Mt, which constituted 4% of the world iron production
(Moore 2001). Yusfin and Pashkov (2002) have projected that the pro-
duction of DRI will reach 53 Mt in 2005.
Presently, most direct reduction processes employ an excess of coal,
which is much higher than the required stoichiometric addition. These
processes are also run at temperatures in excess of 1000 C, which is
higher than necessary for direct reduction to proceed, with the minimum
being around 850 C (Gilchrist 1980).
Although iron ore is the primary source of iron oxide, the existing
reserves need not be utilized at the current rate. Replacing part of the iron
concentrates with recycled steel scrap, iron-containing dust, and metal
machining scrap can reduce the use of natural resources (Heino et al.
2000). In South Africa, Phalaborwa Mining Limited as well as
Foskor Limited produces magnetite as a by-product. The magnetite is
DIRECT REDUCTION OF MIXED MAGNETITE AND COAL PELLETS 65

accumulated in stockpiles consisting of 190 Mt and 35 Mt dumps,

respectively (Minerals Bureau 2001).
Iron ore fed in the form of pellets are a novel type of charge material,
which significantly improve the efficiency of pig iron and steel manu-
facture. In prereduced pellets, the predominant portion of iron oxides is
reduced to metal; in this instance, in the form of sponge iron. Prereduced
pellets consist of a metallized core encased in a partially reduced oxide
shell (Sharma 1997). The direct reduction process is, therefore, greatly
affected by pelletizing, since pellets with a suitable composition are
excellent materials for reduction processes.
Nascimento et al. (1998) have reported that an alternate way of
charging the iron ore and the coal to the reduction reactor is to
agglomerate them together, as pellets. This results in self-reducing pellets.
By using mixed pellets of ore and coal, it is hoped that a lowering of
traditional furnace temperatures, to say 900 C, in addition to allowing
indirect reduction, will initialize direct reduction. This would reduce coal
use (as the product of direct reduction is carbon monoxide, used both as
fuel and in indirect reduction) and should reduce residence time of
material in the furnace.
The use of lignobond, which contains lignosulphonate, as the pellet
binder could have the advantage of increasing carbon levels as compared
to commercial binders, e.g., bentonite. Lignosulphonates are envir-
onmentally friendly by-products of the pulp and paper industry (Lignin
Institute 1991).

PELLET MANUFACTURE AND SUITABILITY

The fines of concern in this study comprise coal as well as magnetite. The
magnetite used was obtained from Martin and Robson Ltd., Middelburg,
South Africa. The coal used was waste coal, normally used as fuel for
boilers, with a low calorific value and high ash content. The coal was
milled in a rod mill and then floated to reduce the ash content.
Two coal size fractions were considered as reductants. Table 1 shows
the specific surface areas and the volume weighted average sizes of the
different feed materials. A Malvern Mastersizer 2000 was used to mea-
sure the size distributions and specific surface areas of the feed materials.
This instrument assumes a homogenous sphere of fixed index of refrac-
tion to determine particle volumes and specific surface areas over a size
distribution (typically accurate between 0.02 and 2000 mm).
66 R. CHELLAN ET AL.

Table 1. Surface areas and sizes of feed material

Magnetite Coarse coal Fine coal

Specific surface area (m2=g) 0.56 0.37 1.44

Volume weighted mean size (mm) 25.85 81.71 14.40

The coarser coal will be referred to as coal <100 mm, while the finer
coal will be referred to as coal <20 mm. The particle sizes of the feed
material are comparable with the material used for pellets in the Fastmet
Process, which are around 40 mm (Fowkes 1997).
The lignobond pellet binder used was obtained from Lignotech SA
and referenced as Lignobond DP001. Carbon present in the binder
amounted to 43.88% (SEM Experimental Analysis). Commercially used
binders, e.g., bentonite, are normally added in the range of 0.5
1% by
mass. A maximum of 1% lignobond binder was therefore added to bind
the pellets. An advantage of lignobond is that it is hydrophilic, thereby
rendering the material to be pelletized wettable. This is advantageous as
coal is naturally hydrophobic.
The pellets required for reduction testing were pelletized mixtures of
magnetite and floated coal fines, including a 1% solid lignobond binder.
The production of pellets was conducted batch-wise in a disc pelletizer,
with 500 g of feed material. The disc pelletizer was operated at 20 revo-
lutions per minute (rpm) at an angle of 60 to the horizontal. The disc
diameter was 61 cm and the lip extended 15 cm from the disc. Pellets were
allowed to dry in an oven at 100 C, but were not heat treated in any way
prior to processing.
It is required that the pellets should not fail under furnace conditions;
therefore, the pellets were subjected to temperatures ranging from
500
900 C and subsequent compression strength tests were performed
using a strength tester. The breaking strengths of both sets of pellets, i.e.,
pellets comprising coal <20 mm and coal <100 mm were found to be greater
than the industrially acceptable limit of 22 N for pellets between 11.2 and
12.7 mm in diameter (Ripke and Kawatra 2000). At temperatures up to
800 C, the pellets made with coarser coal exhibited slightly higher
strength. Pellets showed signs of sintering at temperatures above 800 C
with a significant increase of strength for both types of pellets but a
strength increase of the fine coal pellets to above that for coarse coal
pellets.
DIRECT REDUCTION OF MIXED MAGNETITE AND COAL PELLETS 67

PRELIMINARY DIRECT REDUCTION TESTS

Preliminary experimentation was performed in a tube furnace to observe
if sufficient reduction could occur at 900 C. In each run, a few pellets
were placed in a preheated quartz tube that was then reheated to tem-
perature with an inert gas purge to remove air from the tube. The tem-
perature in the tube was measured with a Pt-100 thermocouple and was
reached in 15 to 20 min. The reactions that were assumed to take place
can be seen below (Gilchrist 1980):

C þ O2 ! CO2 DGo900 C ¼
395 kJ=ðmole O2 Þ
1 ð1Þ

CO2 þ C ! 2CO DGo900 C
35 kJ=ðmole O2 Þ
1 ð2Þ

3Fe2 O3 þ CO ! 2Fe3 O4 þ CO2 DGo900 C ¼
215 kJ=ðmole O2 Þ
1 ð3Þ

Fe3 O4 þ CO ! 3FeO þ CO2 DGo900 C ¼
40 kJ=ðmole O2 Þ
1 ð4Þ

FeO þ CO ! Fe þ CO2 DGo900 C ¼ þ10 kJ=ðmole O2 Þ
1 ð5Þ

3Fe2 O3 þ C ! 2Fe3 O4 þ CO DGo900 C ¼
275 kJ=ðmole O2 Þ
1 ð6Þ

6Fe2 O3 þ C ! 4Fe3 O4 þ CO2 DGo900 C ¼
240 kJ=ðmole O2 Þ
1 ð7Þ

Fe3 O4 þ C ! 3FeO þ CO DGo900 C ¼
100 kJ=ðmole O2 Þ
1 ð8Þ

2Fe3 O4 þ C ! 6FeO þ CO2 DGo900 C ¼
65 kJ=ðmole O2 Þ
1 ð9Þ

FeO þ C ! Fe þ CO DGo900 C ¼
50 kJ=ðmole O2 Þ
1 ð10Þ

2FeO þ C ! 2Fe þ CO2 DGo900 C ¼
15 kJ=ðmole O2 Þ
1 ð11Þ

Reactions (1)
(5) depict indirect reduction due to the carbon mon-

oxide first being formed via the Boudouard reaction, i.e., (1) and (2).
Reported mechanisms for reduction show the sequence of the
reduction process to be hematite (Fe2O3) ! magnetite (Fe3O4) ! wustite
(FeO) ! iron (Fe) (Negri et al. 1987).
68 R. CHELLAN ET AL.

The tests were also intended to investigate the efficiency of using

mixed pellets of iron oxide and waste coal in a furnace. To this end,
mixed pellets comprising magnetite and coal, ratio of 1:1 on a mass basis,
were pelletized inclusive of a 1% binder addition. It should be noted that
the coal addition is in excess to that required for direct reduction. Two
types of pellets were employed, consisting of lignobond binder and
bentonite binder, respectively. Residence times in the tube furnace of 1, 2,
and 3 h were used.
The product pellets were then analyzed by X-ray diffraction (XRD),
which gives relative intensities of hematite, magnetite, wustite, and pure
iron, respectively. Although XRD is a qualitative analysis and not
quantitative, i.e., exact percentages on a mass basis could not be
achieved, it gives an indication of the extent of the reaction. Due to the
presence of hematite, magnetite, wustite, and iron in the reduced pellets,
exact reaction rates could not be evaluated from the products. A com-
parison of relative intensities of materials present from the XRD traces
suggested an increase in iron content over the 3-h period and the high
intensity of the metallic iron peaks gave an indication that sufficient
reduction could occur at the preferred temperature.
The run showing the greatest degree of reduction in the tube furnace
can be seen in Figure 1.

INDUCTION FURNACE TESTS

Experimental Setup
From the tube furnace experimentation it was concluded that reduction of
iron ore to iron could occur efficiently at 900 C in a mixed pellet of waste
coal concentrate and fine magnetite, albeit in a single pellet. Using the same
environment as the tube furnace, i.e., inert atmosphere with a recycle stream
of product gases, the reduction of the pellets needed to be scaled up to
accommodate a batch of pellets. It was evident that the pellet took a long
time to heat up to reduction temperature. Given enough time, i.e., more than
3 h, a greater percentage of iron would have been produced.
An induction furnace would drastically reduce time taken to heat the
pellet as it induces changing magnetic fields within the susceptor, i.e., a
steel casing, which would result in rapid heating of the casing. The casing
would in turn heat the mixed pellets to required temperature. The
induction furnace would also allow for batch heating of an increased
DIRECT REDUCTION OF MIXED MAGNETITE AND COAL PELLETS 69

Figure 1. XRD analysis: mixed ore=floated coal pellets (1:1 by mass), 1% lignobond,
N2 atmosphere, with gas stream.

amount of pellets. Once metallization initiates, the pellets themselves also

act as susceptors.
A schematic of the induction furnace rig is shown in Figure 2. The
induction furnace used was a Mapham Induction Heater, which sup-
plies a maximum output power of 2.5 kW to the heating coils. The

Figure 2. The induction furnace rig.

70 R. CHELLAN ET AL.

cooling water had a maximum temperature of 26 C and a minimum

flowrate of 1.5 l=min, to cool the coils. The glass tube had a high-
temperature refractory lining so that the pellets, while being heated,
would not be in contact with the tube. A distributor plate, which
allowed efficient flow of N2, supported the pellets within the induction
zone, during reduction.
A batch of 500 g of pellets was used in each run. It was found that the
optimum power of 1.2 kW heated the batch of pellets (500 g) to 900 C in
a minimum time of approximately 20 min.
Visual examination of the products showed little observable change
in size or shape. The pellets did, however, exhibit a more metallic luster
and were a lighter gray than before reduction.
After each run, three samples of pellets were analysed by XRD, for
traces of hematite, magnetite, wustite, and iron. Scanning electron
microscope (SEM) analysis was also performed and averaged for the
three samples to acquire amounts of carbon and oxygen remaining.

Results and Discussion

Setting Optimum Parameters. A number of reduction tests had to be
performed in the induction heater, while changing process variables, so
that optimum parameters could be set. Results can be seen in Table 2.
Based on SEM and XRD analysis, the following conclusions were
made.

 A better reducing agent was obtained when a smaller particle size of

coal was used, of the order <20 mm.

Table 2. SEM analysis showing mass percentage of C, O, and Fe in induction heated

pellets under differing conditions

Coal=mag ¼ 1:6 Coal=mag ¼ 1:6 Coal=mag ¼ 1:2 Coal=mag ¼ 1:6 Coal=mag ¼ 1:6
by mass by mass by mass by mass by mass
coal < 100 mm coal < 100 mm coal < 100 mm coal < 20 mm coal < 20 mm
time ¼ 15 min time ¼ 45 min time ¼ 15 min time ¼ 30 min time ¼ 15 min
N2 atmosphere N2 atmosphere N2 atmosphere N2 atmosphere Air atmosphere

C 26.58 23.26 47.48 21.34 18.48

O 17.50 14.09 13.55 12.73 14.82
Fe 52.51 58.93 32.92 58.61 62.15
DIRECT REDUCTION OF MIXED MAGNETITE AND COAL PELLETS 71

 Increasing the coal=magnetite feed ratio, in this case from 1:6 to 1:2 by
mass, i.e., increasing coal addition from 1 to 3 times the stoichiometric
requirement, increases the degree of reduction.
 Reduction in an air atmosphere induces indirect reduction, which is
favorable, but also unfavorably enhances the re-oxidation of reduced
iron.
 Although reduction in an N2 atmosphere nullified re-oxidation of the
product iron, the time taken for a substantial percentage of carbon to
reduce the magnetite was in excess of 45 min.

Optimizing Reduction in Induction Furnace. From the optimization tests,

the following conditions were identified as begin conducive to reduction of
the mixed pellets in the induction furnace. Batches of pellets were prepared
and reduced under these conditions:

 Power input ¼ 1.2 kW (to reach 900 C after 20 min);

 Residence time ¼ 1 h (after 20 min of heating time);
 Magnetite=coal ratio ¼ 2:1 on a mass basis;
 Particle size of coal <20 mm;
 Reduction in nitrogen atmosphere.

XRD and SEM elemental analyses were carried out to identify the degree
of iron formation and the remaining oxide types in the pellets.
The optimized run of duration 1 h showed significant improvement
to the reduction process. The XRD analysis of samples taken at different
sampling points, i.e., top, middle, and bottom (Figure 3), shows traces of
wustite present and no magnetite present. Due to a solid product being
formed, containing large amounts of reduced metallic iron, the product
will now be referred to as DRI.
Using the elemental analysis (Table 3), an approximate mass balance
could be performed to calculate the mass percent of DRI produced.
Assuming that all the oxygen was in the slag phase, the percentage
conversion of magnetite to DRI was calculated to be 100%. This is,
however, improbable, as the XRD analysis show traces of wustite pre-
sent. Assuming that all the oxygen is in the wustite (FeO) phase, mag-
netite is converted to approximately 85% DRI and 15% wustite. This is
also improbable due to oxygen certainly being present in the slag phase.
A single pellet, therefore, comprised approximately 75
80% metallic
iron, 5
15% wustite, 5
10% slag, and 4% carbon. This compares well
72 R. CHELLAN ET AL.

Figure 3. XRD analysis of the optimized run.

with the composition of pig iron, which contains up to 5% carbon

(Peacey and Davenport, 1979).
The residence time of 1 h, however, does not compare well with some
DRI processes, viz. the Fastmet process, which has a residence time of

Table 3. SEM Elemental analyses of pellets

for optimized run

Element Mass Percent

C 3.95
O 3.50
Mg 0.00
Al 2.69
Si 3.31
P 0.00
S 0.33
K 0.00
Ca 1.36
Ti 1.11
Cr 0.00
Mn 0.23
Fe 83.53
DIRECT REDUCTION OF MIXED MAGNETITE AND COAL PELLETS 73

6
12 min (Fowkes 1997). It should be considered, though, that the

Fastmet mixed pellets contains 5 times more coal than the stoichiometric
amount required, while the current process contains a reduced amount of
3 times stoichiometric requirement.
Energy consumption of the process amounted to a total of
33.65 GJ=ton DRI, with induction heating being 9.94 GJ=ton DRI (based
on the input power to the induction furnace and the heating time) and
energy consumption of coal being 23.71 GJ=ton DRI (based on the
amount of coal used as compared to metal produced with a calorific value
for the domestic coal used). The blast furnace calorific consumption
amounts to approximately 17.9 GJ=ton hot metal. This value is based
only on coke consumption and does not consider energy consumed for
heating the furnace to around 1800 C, i.e., almost double the tempera-
ture of the current process. Even though the coal consumption maybe
higher, coke is more expensive and in short supply as compared to waste
coal. The calorific consumptions of coal-based processes based only on
coal consumption for production of DRI range between 15 and
25 GJ=ton DRI.

Advantages of the DRI Process

The general advantages of using an induction heater are high energy, high
efficiency, rapid heating, cleanliness, and process repeatability (Khan
et al. 2000). The use of the induction heater in the current invention
reduces amount of coal fed in, as coal is not needed as a fuel, but solely as
a reducing agent.
Due to the operating temperature being only 900 C, excessive
quantities of high-temperature refractory material are not needed as in
blast furnace processes, i.e., processes run in excess of 1400 C. This
would reduce the cost of refractory material to a great extent.
Molten reactions are more corrosive to refractories than solid state
reduction (as there is increased contact). Another advantage of the low
temperature is that it prevents the formation of molten products, which
may be lost to refractories and could cause problems for induction
heating.
If the current invention was to be scaled up for industrial applica-
tions, the capacity of the induction heater could be increased, by man-
ufacturers, to handle industrial amounts of feed, i.e., power supplies with
output power ranging from 500 W to 2500 kW at frequencies ranging
74 R. CHELLAN ET AL.

from 10 kHz to 30 MHz are available from manufacturers (Anon 2003).

Alternate susceptor material would, however, have to be investigated, as
there may be a danger of partial melting of the susceptor while the pellets
are heating when using larger batches.

Uses of the DRI Process

Due to the product being a solid DRI of purity >75%, further processing
is obviously required to produce a molten hot metal such as the blast
furnace product pig iron. This process, in effect, produces a prereduced
pellet.
This process could, therefore, be used as prereduction step before
final reduction and melting of the finished product. The Corex (Jones
2001) and Fastmet processes (Tanaka et al. 2002) both have similar
prereduction steps, albeit reduction of the pellets occurs in a shaft and
rotary hearth furnace respectively. In fact, reduction via induction
heating, if proven on an industrial scale, could provide a more easily
controlled lower temperature process.

CONCLUSIONS
A new DRI process using induction heating was investigated and the
following conclusions were made.

 Reduction occurs quite effectively at 900 C, although indirect reduc-

tion plays the major role. Haque and Ray (1995) also arrived at the
same conclusion that one cannot prevent indirect reduction.
 Lignobond is an effective binder for low-temperature processes and
could be used as a replacement for bentonite. It also adds carbon to the
process and removes some slag forming minerals.
 Preliminary testing of the pellets in a tube furnace demonstrated that
floating of the waste coal improved the reducing capabilities of the
coal. Recycling the product gases, i.e., CO and CO2, ensures that CO
reduces unreacted iron oxide indirectly, thereby increasing the con-
version of magnetite.
 An optimized run in the induction heater, upon analysis by XRD and
SEM, produced pellets comprising approximately 75
80% metallic
iron, 5
15% wustite, 5
10% slag, and 4% carbon.
DIRECT REDUCTION OF MIXED MAGNETITE AND COAL PELLETS 75

 Energy consumption of the process amounted to a total of 33.65 GJ=ton

DRI, with induction heating being 9.94 GJ=ton DRI and energy con-
sumption of coal being 23.71 GJ ton DRI. The energy efficiency was
found to be favorable as compared to some DRI processes.

REFERENCES
Anon, 2001, ‘‘Preliminary energy outlook for South Africa,’’ Energy Research
Institute, Dept. of Mechanical Engineering, University of Cape Town, South
Africa, Rondebosch, www.dme.gov.za=energy=iep.pdf
Anon, 2003, ‘‘Induction heating,’’ Lepel, http:==inductionheating.com
Els, J., 2001, Next Generation Developments in the Ferrous Industry:
Pyrometallurgy: Current Issues and Future Trends, Mintek, Randburg, The
South African Institute of Mining and Metallugy.
Feinman, J., 1999, Direct Reduction and Smelting Process, ch. 11, Pittsburgh, PA,
AISE Steel Foundation, J. Feinman and Associates, Inc., www.aise.org=
newpubs=ironchap%2011.pdf
Fowkes, N., 1997, ‘‘The reduction of iron ore pellets in the Fastmet process,’’
Research Report 1994=31. Mathematics Department, University of Western
Australia, Australia.
Gielen, D., 2002, CO2 Removal in the Iron and Steel Industry, Ibarahi, Japan,
National Institute for Environmental Studies.
Gilchrist, J. D., 1980, Extraction Metallurgy, 2nd ed., London, U.K., Pergamon.
Haque, R. and Ray, H. S., 1995, ‘‘Role of ore=carbon contact and direct
reduction in the reduction of iron oxide by carbon,’’ Mettallurg. Materials
Trans. B, 26B, pp. 400 401.
Heino, J., Makkonen, H., Laitila, L., Hatkki, J., and Poylio, E., 2000, Recycling
or utilisation of dust, scales and sludge from steel industry, www.cc.jyu.fi=
helsie=pdf=heino.pdf
Jones, R. T., 2001, Iron and Steel, Murdoch University, Perth, Western Australia,
Australia, www.science.murdoch.edu.au
Khan, S., Ahamed, M., Khan, M. A., and ul-Haq, A., 2000, ‘‘Innovative appli-
cations of induction heating for selective heat treatment,’’ Sci. Vision Quar-
terly, 5(4), pp. 14 19.
Lignin Institute, 1991, What is Lignin? www.lignin.info
Lungen, H. B., 2001, ‘‘State of the art and future of the blast furnace,’’ Paper
presented at the 35th Meeting of the IPIS, Berlin, Germany, www.pig-iron.
com=Informationen=images=Luengen.pdf
Minerals Bureau, 2001, Producers of Iron Ore, Directory D8=2001, www.dme.gov.za=
publications=pdf=project_research=minerals=d8_2001=d8_2001 02.pdf
76 R. CHELLAN ET AL.

Moore, C., 2001, ‘‘120 years in pig iron,’’ Proc. 60th Electric Furnace Conf., Int.
Pig Iron Secretariat, Dusseldorf, Germany.
Nascimento, R. C., Mourao, M. B., and Capocchi, J. D. T., 1998, ‘‘Reduction-
swelling behaviour of pellets bearing iron ore and charcoal,’’ Canadian
Metallurg. Quarterly,. 37(5), pp. 441 448.
Negri, E. D., Alfano, O. M., and Chiovetta, M. G., 1987, ‘‘Direct reduction of
hematite in a moving bed. Comparison between one- and three-interface
pellet models,’’ Chem. Eng. Sci., 42(10), pp. 2472 2475.
Peacey, J. G. and Davenport, W. G., 1979, The Iron Blast Furnace — Theory and
Practice, London, U.K., Pergamon.
Ripke, S. J. and Kawatra, S. K., 2000, ‘‘Can fly-ash extend bentonite binder for
iron ore agglomeration?,’’ Int. J. Miner. Process., 60, pp. 181 198.
Sasaki, N., 1998, Direct Iron-Making Process Using Fusion Reduction (DIOS),
ICETT, Japan, www.icett.or.jp=techinfo.nsf
Sharma, T., 1997, ‘‘Reduction of double layered iron ore pellets,’’ Int. J. Miner.
Process., 49, pp. 201 206.
Tanaka, H., Sugiyama, T., Harada, T., and Sugitatsu, H., 2002, FASTMET Dust
Pellet Reduction, Operations report on the First Fastmet Waste Recovery
Plant, Kobe Steel, Ltd. Japan.
Yusfin, Y. and Pashkov, N., 2002, ‘‘Non-blast furnace steel: Alternative or
supplement?,’’ Eurasian Metals, National Review, en.eurasmet.ru=online=
2002=1=science&tech.html